Supported Providers

Getting your API key: 1. Go to [platform.openai.com/api-keys](https://platform.openai.com/api-keys). 2. Click Create new secret key. 3. Copy the key (starts with sk-).

Recommended models: - gpt-4o -- Best balance of quality and speed (128K context). - gpt-4o-mini -- Faster and cheaper, good for autocomplete and simple tasks. - o3 -- Reasoning model for complex problem-solving. Does not support temperature parameter. - gpt-5 -- Latest flagship model with native tool calling.

Notes: - OpenAI reasoning models (o1, o3, o4 series) do not support the temperature parameter. QCoder automatically detects this and retries without temperature. - GPT-5 supports native tool calling and is not classified as a reasoning model.

Getting your API key: 1. Go to [console.anthropic.com/settings/keys](https://console.anthropic.com/settings/keys). 2. Click Create Key. 3. Copy the key (starts with sk-ant-).

Recommended models: - claude-sonnet-4-20250514 -- Excellent coding model with 200K context. - claude-opus-4-20250514 -- Most capable model for complex tasks.

Notes: - Anthropic models have a 200K token context window, the largest among commercial providers. - Native tool calling is fully supported.

Getting your API key: 1. Go to [aistudio.google.com/apikey](https://aistudio.google.com/apikey). 2. Click Create API key. 3. Copy the key (starts with AIza).

Recommended models: - gemini-2.0-flash -- Fast and capable, good for most tasks. - gemini-2.5-pro -- Most capable Gemini model with 1M+ context window.

Notes: - Gemini models support the largest context windows available (1M+ tokens), making them excellent for working with large codebases. - Native tool calling is supported.

Getting your API key: 1. Go to [platform.deepseek.com](https://platform.deepseek.com). 2. Navigate to API Keys and create a new key. 3. Copy the key (starts with sk-).

Recommended models: - deepseek-chat (V3) -- Strong coding model at very competitive pricing. - deepseek-reasoner -- Reasoning model for complex problem-solving.

Notes: - DeepSeek offers some of the best price-to-performance ratios for coding tasks. - The API is OpenAI-compatible, so it works seamlessly with QCoder's OpenAI integration.

Getting your API key: 1. Go to [console.groq.com/keys](https://console.groq.com/keys). 2. Create a new API key. 3. Copy the key (starts with gsk_).

Recommended models: - llama-3.3-70b-versatile -- Fast, high-quality general coding. - mixtral-8x7b-32768 -- Good for autocomplete due to very fast inference.

Notes: - Groq specializes in ultra-fast inference using custom LPU hardware. - Ideal for inline completions and autocomplete where low latency matters.

Getting your API key: 1. Go to [console.mistral.ai/api-keys](https://console.mistral.ai/api-keys). 2. Create a new key.

Recommended models: - codestral-latest -- Purpose-built for code generation. - mistral-large-latest -- General-purpose, strong reasoning.

Notes: - Mistral has a 120-second request timeout, which is longer than most providers. QCoder accounts for this automatically. - European-hosted, which may be relevant for data residency requirements.

Both Ollama and LM Studio are free local providers that run models on your own hardware. No API key is needed and no data ever leaves your machine.

Ollama runs as a background service and is controlled from the command line. It is the most popular local model runner.

LM Studio provides a graphical interface for downloading, managing, and serving models. It is ideal if you prefer a visual workflow.

Both providers communicate via an OpenAI-compatible local API, so QCoder treats them the same as cloud providers -- just without authentication.

For local providers, tool calling is handled via XML prompting rather than native function calling, since most local models do not support the OpenAI function calling format.